In an article in today's Tax Notes (here, but it's behind a paywall), DC tax lawyers Donald Susswein, Kyle Brown, and J. Anthony Coughlan report their "surprising finding that there are billions in deferred corporate and individual capital gains taxes missing from the official tax expenditure budgets."
What they have in mind is realized but not recognized capital gains from tax-free corporate mergers and acquisitions, in cases where the transactions aren't mere changes in legal form (like reincorporating an existing corporation with all the same owners for some sort of minor technical reason). Rather:
"We are talking about exchanges that substantially change the parties’ economic holdings before and after the deal. Examples would include the tax-free merger of companies like and into or the tax-free transfer of $1.65 billion in stock to the founders of a 2-year-old startup like , deferring individual and corporate capital gains taxes on as much as $1.65 billion of realized capital gains in a single deal. These and thousands of similar arm’s-length exchanges clearly meet the official definition of a tax expenditure (there would be no feasibility problem with taxing the or deals if they failed to satisfy all of the technical requirements for tax-free treatment). But they are seemingly invisible to the tax expenditure budgets."
Their point is not that the underlying recognition rule is bad policy (reflecting that classifying something as a tax expenditure does not necessarily mean that it is bad policy). Rather, it's that these are clearly "income" under the "normal" income tax rules - which only count realized, not unrealized capital gains - on which the tax expenditure budget has long relied.
The article takes a bit of a wrong turn in relating all this to purported ambiguities in the Haig-Simons income definition. What makes that a wrong turn is the fact that the conventional tax expenditure budget expressly does not rely on the Haig-Simons concept, which again would apply even to unrealized gains.
I think that their article nonetheless makes an interesting point that may advance one's thinking about the use of tax expenditures as a concept. If one were to disagree with it within the conventional definition of tax expenditures (relying on the "normal" income tax structure, one would say that this structure excludes, not only unrealized gains, but what one might call (for want of a better term), kind-of, almost, quasi-unrealized gains. They themselves say that "[n]o one could object to not listing [mere changes in legal form] as tax expenditures." So they aren't actually drawing the line, for "normal" structure purposes, at technical realization vs. no technical realization. Rather, they are drawing it at technical realization plus something of economic significance changes as a result of the transaction.
This is a perfectly respectable way to define the "normal income tax structure" for tax expenditure purposes. Only, raising it highlights the unclarity of the underlying question: Why are we using that concept to begin with? What normative arguments should govern the debate about that choice?
When I wrote about tax expenditures some years ago - indirectly inspiring, I gather, the Joint Committee of Taxation's short-lived efforts to revise the concept when Ed Kleinbard was the JCT Chief of Staff - I thought that a far more coherent and useful concept than normal versus not normal structure was the famous distinction in Richard Musgrave's public economics work between the government's distributional and allocative branches. E.g., the distributional branch might or might not want to allow deductions for medical expenses (a long-debated policy question). But it definitely would not distinguish between municipal and corporate bonds, which both income and consumption tax advocates agree should be taxed the same (whether both are included or both excluded) if one is merely trying to measure distribution without regard to implicit taxes. And it also would not countenance a tax credit or deduction for solar heating panels, as only the allocative branch would consider that a possibly worthwhile rule. I argued that, in the ordinary case, what we call a "tax expenditure" is an allocative rule (whether good or bad) that has been placed in what we think of as a mainly distributional system.
I thought that an improved version of the concept would distinguish between (a) rules that were clearly just allocative (at least, when thought about in a principled way), (b) rules that were clearly within the distribution branch's (such as deducting losses and gains in order to calculate net income), and (c) those as to which the distribution branch's preferred treatment was legitimately debatable (such as medical deductions)
[Let me insert here a brief detour. I have completed writing and will soon be posting on SSRN an article about medical deductions et al, which discusses among other things the recently much debated issue of deductions for the cost of assisted reproductive technologies or ART.]
Okay, back to the main topic. I thought that there should also be a category in the revised tax expenditure budget for items that reflect (at least in part) the concerns of the distribution system but that respond to administrative concerns rather than the aim of directly & correctly measuring (say) Haig-Simons income. This is where I would have put unrealized gains. It would be good, if the TE budgets are at all relevantly informative, to include a category for unrealized gains even though they are not, at least exclusively or even primarily, an intervention by the allocation system for incentive or other such reasons.
So the Susswein et al point could be construed as holding that amounts realized but not recognized by reason of the nonrecognition rules for corporate transactions should be divided into 2 categories. Those that apply to mere changes in legal form would go into my category of rules reflecting the administrative et al concerns of the distribution system. Those for the likes of the Exxon-Mobil and Google-YouTube transactions would be classified as tax expenditures. Not an unreasonable idea at all, although one would need to come up with and apply a dividing line.
Then again, maybe all this (including my work as well as theirs) vastly exceeds the real policy gains, if any, to be derived from trying to rationalize the tax expenditure budget. But I think there is some intellectual value to trying to get the underlying concepts (more) straight, and the Susswein et al article is helpful in this regard.
But one last quibble: Susswein et al argue that clarifying these points, while "obviously relevant to the tax expenditure budgets ... may be even more important to the question of whether unrecognized gains should be included in the determination of average tax rates (and other tax policy questions)." Here I disagree. When we are thinking, say, about the overall distributional effects of the tax system, or the broader fiscal system - e.g., how favorably does it treat the super-rich relative to those below them - Haig-Simons income is one of the distributional measures that one might consider using. (As I discuss here, another possibility is lifetime income.) The "normal" income tax structure really has no role to play in, and associated questions of what set of underlying motivations particular income tax rules might address, has really no role to play in, and nothing to contribute to, that discussion.